Book an intro call
All case studies
Carb Estimator · Digital health

Automating the complex carb counting process for people with diabetes.

The Carb Estimator app on a phone next to a plate of food

Challenge

Help people with diabetes better estimate the carbs in their meals.

People with diabetes, specifically Type 2, struggle to estimate the carbohydrates in their meals. This is important as patients can experience hypoglycaemia through incorrectly estimated carb intake and insulin bolus.

  • People with diabetes often have to spend a lot of time calculating their carb intake and adjusting their insulin doses.
  • This can be a burden, especially for people with busy lives.
  • Ineffective management of blood glucose can lead to other chronic health complications, such as weight gain, fatigue, and difficulty sleeping.

Approach

  1. Research

    Research with diabetes patients showed a need for a better way to count carbohydrates, as using tables is time-consuming and cumbersome.

  2. Design

    We conducted a user-centered design process creating the UX and UI iteratively with real user input.

  3. Develop

    We developed an AI-driven solution using cross-platform machine learning algorithms for Android and iOS.

Solution

A smartphone app allowing patients to get carbohydrate value from a photo of their meal.

  • App screen recognising the food items in a photo of a meal

    AI-augmented image recognition for documenting meals.

    People use the app to take a photo of their meal, which utilises image recognition technology to identify the meal itself, and the individual food items within it.

  • App screen showing the estimated carbohydrates for each food item

    Algorithmically estimate carbohydrate count automatically.

    AI and machine learning algorithms estimate the carbohydrate content of each food item and produce an estimated total carb value for the meal, greatly reducing the burdensome task of carb counting for the user.

  • App screen recommending an insulin bolus dose for the meal

    Personalised bolus dose recommendation.

    Combining data from the meal’s carbohydrate count, patient biometrics and their inputted basal profile, the app can then provide a personalised insulin bolus dose for that meal.

Outcome

8 out of 10 users said they would continue using the app after the pilot.

  • For patients

    Gives patients more confidence in carb values when choosing meals, allowing them to spend less time worrying about tedious carb counting.

  • For HCPs

    Allowing HCPs to make better, more informed therapy decisions, supported by a detailed log of carb intake data alongside typical data points of CGM and basal/bolus doses.

  • For payers

    Reduced co-morbid health complications by giving patients better self-management tools.

More case studies

All case studies

Ready to talk?

Book a 30-minute introductory call to explore how our expertise can drive innovation in your healthcare product.

Book an intro call Or email us at